SpringerLink
Forum Springer Astron. Astrophys.
Forum Whats New Search Orders


Astron. Astrophys. 353, L13-L17 (2000)

Previous Section Next Section Title Page Table of Contents

2. Observations and data reduction

NGC 6946 was observed at 450 µm and 850 µm, during April 10, 11 and June 17, 18, 19, 20 1998.

SCUBA consists of two bolometer arrays of 91 elements optimised to observe at 450 µm and 37 elements optimised at 850 µm, covering a field of view of about 2.3 arcmin (Holland et al. 1999). The camera, mounted on the Nasmyth focus of the telescope can be used simultaneously at both wavelengths, by means of a dichroic beamsplitter.

In the scan-map mode, the telescope scans the source at a rate of 24 arcsec per second, along specific angles to ensure a fully sampled map. Meanwhile the secondary chops with a frequency of 7.8 Hz within the observed field. While this ensures a correct subtraction of the sky background, the resulting maps unfortunately have the profile of the source convolved with the chop. The profile of the source is restored deconvolving the chop from the observed map by mean of Fourier Transform (FT) analysis.

Scan-maps of NGC 6946 presented here are fully sampled over an area of 8´x8´. Each set of observations consisted of six scans, with different chop configurations: chop throws of 20", 30" and 65" along RA and Dec are needed to retrieve the final image. Data have been reduced using the STARLINK package SURF (Jenness & Lightfoot 1999). Images were first flat-fielded to correct for different sensitivities of the bolometers. Noisy bolometers were masked and spikes from transient detections removed by applying a 5-[FORMULA] clip. A correction for atmospheric extinction was applied, using measures of the atmosphere opacities taken several times during the nights of observation. Zenith optical depth varied during the six nights, with [FORMULA] and [FORMULA]. The 450 µm opacity on the last night was too high ([FORMULA]) for the source to be detected and therefore the relative maps were not used for this wavelength. Because of the chopping in the source field, each bolometer sees a different background: a baseline, estimated from a linear interpolation at the edges of the scan, has been subtracted from each bolometer.

Sky fluctuations were derived from the time sequence of observations for each bolometer, after the subtraction of a model of the source, obtained from the data themselves. The images have then been corrected by subtacting the systematic sky variations from each bolometer.

Data taken with the same chop configuration were rebinned together into a map in an equatorial coordinate frame, to increase the signal to noise. Six maps with 3" pixels were finally obtained for each wavelength, combining 33 and 25 observations, at 850 and 450 µm, respectively. In each of the six maps the signal from the source is convolved with a different chop function. The final deconvolved image is retrieved using the Emerson II technique (Holland et al. 1999; Jenness, Lightfoot & Holland 1998). Essentially, for each rebinned image, the FT of the source is derived by dividing the FT of the map by the FT of the chop function, a simple sine-wave. Since the division boosts up the noise near the zeros of the sine-wave, different chop configuration are used. For the chosen chop throws, the FT of the chop functions do not have coincident zeros, apart from the zero frequency. A smoother FT of source can therefore be obtained, and the final image is retrieved by the applying an inverse FT.

Unfortunately the deconvolution introduces artifacts in the images, like a curved sky background. This may be due to residual, uncorrected, sky fluctuation at frequencies close to zero, where all the chops FT goes to zero. Work to solve this problem is ongoing (Jenness, private communication). To enhance the contrast between the sky and the source, we have modelled a curved surface from the images, masking all the regions were the signal was evidently coming from the galaxy. The surface has been then subtracted from the image.

Calibration was achieved from scan-maps of Uranus, that were reduced in the same way as the galaxy. Integrated flux densities of Uranus were derived, for each observing period, using the STARLINK package FLUXES (Privett, Jenness & Matthews 1998) for JCMT planetary fluxes. Comparing data for each night we derived a relative error in calibration of 8 per cent and 17 per cent, for 850 µm and 450 µm respectively. From the planet profile, the beam size was estimated: FWHMs of 15.2" and 8.7" were measured for the beam at 850 and 450 µm, respectively. To increase the signal to noise, the 850 µm image has been smoothed with a gaussian of 9" thus degrading the beam to a FWHM of 17.7" The 450 µm image has been smoothed to the same resolution as for the 850 µm one, to facilitate the comparison between features present in both. The sky [FORMULA] in the smoothed images is 3.3 mJy beam-1 at 850 µm and 22 mJy beam-1 at 450 µm.

The final images, after removing the curved background and smoothing are presented in Fig. 1. For each wavelength, the grey scale shows all the features [FORMULA]1-[FORMULA], while contours starts at 3-[FORMULA] and have steps of 3-[FORMULA].

[FIGURE] Fig. 1. Sub-mm images of NGC 6946, at 850 µm (top-left) and 450 µm (top-right). Grey scales show features 1-[FORMULA] above the sky, while contours starts at 3-[FORMULA] and have steps of 3-[FORMULA]. Both images have a beam size FWHM=17.7" An area of 10´x10´ is displayed, but only the central 8´x8´ are fully sampled. North is on top, East on the left. A U-band image of NGC 6946 (Trewhella 1998) and a 12CO(2-1) emission map (Sauty, Gerin & Casoli 1998), are presented in the bottom-left and bottom-right panels, respectively. 850 µm contours are overlayed to both optical and line emission images. The centre of the galaxy is in RA=20h 34m [FORMULA] and Dec=60o 09´ [FORMULA] (J2000; De Vaucouleurs et al. 1991). The scale is 27 pc arcsec-1 (D= 5.5 Mpc; Tully 1988).

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: December 17, 1999
helpdesk.link@springer.de